Here is a 'KRRBN vs krrp' four-move chess problem generated by a computer using the Digital Synaptic Neural Substrate (DSNS) computational creativity approach. It doesn't use endgame tablebases, neural networks or any kind of machine learning found in traditional AI. There is also no proven limit to the quantity or type of legal compositions that can be automatically generated. This position contains a total of 9 pieces. The largest endgame tablebase in existence today is for 7 pieces (containing over 500 trillion positions anyway) which means the problem could not have been taken from it regardless.
White to Play and Mate in 4
Chesthetica v10.82 : Selangor, Malaysia
2018.10.20 4:17:43 PM
Some of the earliest chess problems by humans are over 10 centuries old but original ones by computer are very recent. Chesthetica composes only unique or new constructs. If you have seen it before, cite the source and comment below because it is purely coincidental. Do share and try out some of the others too. Collectively, these puzzles are intended to cater to players of all levels. If you're bored of standard chess, though, why not try this?
Main Line of the Solution (Skip to 0:35)
<| Book | DTube | Facebook | Reddit | Twitter | Website |>